Chemical Physics Letters 689 (2017) 26–29 Contents lists available at ScienceDirect Chemical Physics Letters journal homepage: www.elsevier.com/locate/cplett Research paper High heating rate decomposition dynamics of copper oxide by nanocalorimetry-coupled time-of-flight mass spectrometry ⇑ ⇑ Feng Yi a, Jeffery B. DeLisio b, Nam Nguyen a, Michael R. Zachariah b, , David A. LaVan a, a Materials Measurement Science Division, Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, United States b Department of Chemistry and Biochemistry, and Department of Chemical and Biomolecular Engineering, University of Maryland, College Park, MD 20742, United States article info abstract Article history: The thermodynamics and evolved gases were measured during the rapid decomposition of copper oxide Received 18 July 2017 (CuO) thin film at rates exceeding 100,000 K/s. CuO decomposes to release oxygen when heated and In final form 28 September 2017 serves as an oxidizer in reactive composites and chemical looping combustion. Other instruments have Available online 30 September 2017 shown either one or two decomposition steps during heating. We have confirmed that CuO decomposes by two steps at both slower and higher heating rates. The decomposition path influences the reaction course in reactive Al/CuO/Al composites, and full understanding is important in designing reactive mix- tures and other new reactive materials. Ó 2017 Published by Elsevier B.V. 1. Introduction As will be described below, CuO films were deposited directly onto the nanocalorimeter by sputtering from a CuO target with a Accurate measurements of the thermal decomposition of CuO thin layer of alumina as a diffusion barrier between the silicon at high heating rates is directly related to developing a deeper nitride sensor membrane and the sample. X-ray diffraction was understanding of energetic composites [1,2], chemical looping used to confirm the deposited phase and valence state before start- combustion [3,4] and solar thermochemical fuel production [5]. ing the experiments. The nanocalorimeter-TOF MS instrument was In contrast, CuO is also widely used as a catalyst, but catalytic used to measure the decomposition process and capture decompo- applications are primarily interested in avoiding the decomposi- sition temperatures and evolved species information. tion of CuO [6]. In traditional slow heating rate studies using thermogravimetric analysis and differential scanning calorimetry 2. Method and experiments (TGA/DSC), CuO decomposes to Cu2O, then to Cu in a two-step decomposition process. However, in a recent article [7], only 2.1. Sample preparation one oxygen evolution event was observed when CuO was rapidly  5  5 heated at heating rates between 1.5 10 K/s and 6.5 10 K/s, The CuO and Al films were prepared using sputter deposition which suggested the possibility of a different decomposition with a copper backed CuO target (Kurt Lesker, 99.7% purity) and mechanism: two decomposition steps for low heating rates and an Al target (Kurt Lesker, 99.9995% purity) under a pressure of 5 one decomposition step for high heating rates; we have investi- mTorr of argon. RF power of 300 W was used for CuO sputtering gated this materials system further with our new instrument to and DC power of 300 W was used for aluminum sputtering. The provide more detailed data. nanocalorimeter sensor fabrication and calibration have been Nanocalorimeters are MEMS-based thermal sensors capable of described previously [13]. It has a 100 nm thick platinum heater measuring thermodynamic properties of very small amounts of suspended on a 100 nm thick silicon nitride membrane in a silicon sample with high sensitivity under very fast heating rates [8–11] frame. CuO films are deposited on the silicon nitride side of the 6 up to 10 K/s. Recently, a newly developed instrument integrating nanocalorimeter sensor with a thin alumina barrier layer between a nanocalorimeter into a time-of-flight mass spectrometer (ToF- the sample and sensor. MS) has been demonstrated to simultaneously measure thermal properties and evolved gas phase species of rapid reactions [12]. 2.2. Nanocalorimetry-coupled time-of-flight mass spectrometry measurements ⇑ Corresponding authors. E-mail addresses: [email protected] (M.R. Zachariah), [email protected] The integration of the nanocalorimeter into a time-of-flight (D.A. LaVan). mass spectrometer (ToF-MS) has recently been described [12]. https://doi.org/10.1016/j.cplett.2017.09.066 0009-2614/Ó 2017 Published by Elsevier B.V. F. Yi et al. / Chemical Physics Letters 689 (2017) 26–29 27 Briefly, the nanocalorimeter sensor was positioned at the ioniza- different mechanisms comparing faster heating rates to slow tion region of the ToF-MS using a linear motion feedthrough with heating rates. a 3D printed adapter. These reactions were typically completed To prevent the possible reaction between oxygen released within 10 ms. The ToF-MS was sampled at 100 ms per spectrum from the sample and the silicon nitride sensor membrane at (10 kHz), with 100 spectra obtained post-triggering for each run. temperatures over 1000 °C [17], a 40 nm barrier layer of alu- Gas phase reaction products were ionized for 3 ls using an elec- mina is deposited between sample and silicon nitride. tron gun operated at 70 eV and 1 mA. Nanocalorimetry results shown in Fig. 2 provide the first insights into these materials – the peak expected for Cu4O3 3. Results and discussion decomposition to CuO and Cu2O would have occurred in a broad exothermic reaction with a peak temperature at 427 ° 3.1. Copper (II) Oxide phase confirmation C [14], but this peak is absent, which supports the x-ray diffraction data showing the deposited material is CuO instead of Cu4O3. Copper is multivalent and can form CuO, Cu4O3 and Cu2O. Typ- As shown in Fig. 2a, two endothermic signals are observed with ically, copper oxide films are deposited by sputtering from either ° ° Cu or CuO targets, and their final chemical form would depend peak temperatures of 975 C and 1112 C respectively, indicating on the oxygen availability and other sputtering parameters. In this two step decompositions of CuO. However, as is true for the previ- case, the target material is CuO and no reactive gas (oxygen) is ous mass spectrometer results, there is only one broad oxygen evo- introduced. Fig. 1 shows the X-ray diffractogram of the deposited lution event that spans the temperature range of the two film. The patterns and relative intensities of the diffraction peaks endothermic thermal signals. Therefore, identifying this reaction match with JCPDS data for CuO (PDF 00-048-1548). The main as one step or two steps based solely on mass spectrometer data diffraction peaks of 35.3° and 38.3° correspond to the lattice plane depends on the rate of removal of oxygen from the system and of (À1 1 1) and (1 1 1) respectively. Therefore, the result here con- the temporal resolution of the mass spectrometer. We see two firmed the sputtered film is CuO. A previous article [14] reported peaks in the nanocalorimeter data due to this instrument’s greater temporal resolution. that a Cu4O3 film was formed while sputtering from a CuO target, but we did not see this in our samples. It is likely due to differences Efforts were made to further separate the oxygen release in the sputtering tool or parameters. First, the base pressure in that events. The heating and measurement duration was varied from report was 1.9  10À7 Torr, lower than in this work (3  10À6 torr), 10 ms to 50 ms. As shown in Fig. 2b, two consecutive endothermic therefore more oxygen was available for our samples. Second, that signals were observed accompanied by only one oxygen release substrate was water-cooled but our substrate was not temperature event in the TOF-MS data. The oxygen signal spans the two controlled. endothermic peaks even though the time between the two signals is increased to 2 ms from less than 1 ms. Here, we also see the tradeoff between the mass spectrum signal sensitivity and heating 3.2. Decomposition of CuO film rate. The gap between the two decomposition events is only a few milliseconds in the thermal data and yet is not resolved in the mass The decomposition of CuO is believed to occur in two steps: spectrometer signal. CuO first decomposes to Cu O and O , followed by the decompo- 2 2 To further study the decomposition of CuO, a sample was sition of Cu O to Cu and O at a higher temperature. This process 2 2 rapidly heated to a maximum temperature below the onset has been reported previously [15,16] and measured at slow rates of the second endotherm. These results are shown in Fig. 3, using differential scanning calorimetry of CuO nanoparticles (data the sample was heated to 1005 °C, which is just below the not shown). However, a recent article[7] suggests only one onset temperature of the second decomposition step. No nitro- oxygen release event during rapid heating (from 1.5  105 K/s gen was detected (meaning the barrier layer prevented a reac- to 6.5  105 K/s) based on observations in a mass spectrometer. tion with the membrane) and no exothermic signal is A change in the number of peaks suggests kinetic limitations or observed. The only oxygen source is from the first decomposi- tion step and the peak temperature of oxygen release is at 982 °C, which is lower than the peak oxygen release tempera- ture (1070 °C) for the sample heated up to 1270 °C, as shown in Table 1. The results from the sample heated to 1270 °C cov- ers the temperature range of both the decomposition steps and shows the oxygen is from both decomposition steps.